Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -37,9 +37,10 @@ summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
|
| 37 |
|
| 38 |
|
| 39 |
|
|
|
|
| 40 |
class SmartAgentV2:
|
| 41 |
def __init__(self):
|
| 42 |
-
self.qa_model = pipeline("text2text-generation", model="google/flan-t5-
|
| 43 |
self.summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
| 44 |
self.whisper_model = whisper.load_model("base")
|
| 45 |
|
|
@@ -74,6 +75,15 @@ class SmartAgentV2:
|
|
| 74 |
result = self.whisper_model.transcribe(filepath)
|
| 75 |
return result["text"]
|
| 76 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 77 |
def sum_food_sales(self, filepath):
|
| 78 |
df = pd.read_excel(filepath)
|
| 79 |
food_df = df[df["Category"].str.lower() == "food"]
|
|
@@ -94,6 +104,10 @@ class SmartAgentV2:
|
|
| 94 |
if filepath := getattr(file, "name", None):
|
| 95 |
if filepath.endswith(".mp3"):
|
| 96 |
transcript = self.transcribe_audio(filepath)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
return transcript
|
| 98 |
elif filepath.endswith(".xlsx") or filepath.endswith(".xls"):
|
| 99 |
return self.sum_food_sales(filepath)
|
|
@@ -113,6 +127,7 @@ class SmartAgentV2:
|
|
| 113 |
return "An error occurred processing the question."
|
| 114 |
|
| 115 |
|
|
|
|
| 116 |
def run_and_submit_all( profile: gr.OAuthProfile | None):
|
| 117 |
"""
|
| 118 |
Fetches all questions, runs the BasicAgent on them, submits all answers,
|
|
|
|
| 37 |
|
| 38 |
|
| 39 |
|
| 40 |
+
|
| 41 |
class SmartAgentV2:
|
| 42 |
def __init__(self):
|
| 43 |
+
self.qa_model = pipeline("text2text-generation", model="google/flan-t5-xl")
|
| 44 |
self.summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
| 45 |
self.whisper_model = whisper.load_model("base")
|
| 46 |
|
|
|
|
| 75 |
result = self.whisper_model.transcribe(filepath)
|
| 76 |
return result["text"]
|
| 77 |
|
| 78 |
+
def extract_ingredients(self, transcript):
|
| 79 |
+
ingredients = re.findall(r"(?:\ba|\ban|\bthe)?\s*([a-zA-Z\s]+?)\s*(?:\bof\b|\bcups?\b|\btablespoons?\b|\bteaspoons?\b|\bpinch\b)?", transcript)
|
| 80 |
+
ingredients = [i.strip().lower() for i in ingredients if len(i.strip()) > 2]
|
| 81 |
+
return ", ".join(sorted(set(ingredients)))
|
| 82 |
+
|
| 83 |
+
def extract_page_numbers(self, transcript):
|
| 84 |
+
numbers = re.findall(r"\b\d+\b", transcript)
|
| 85 |
+
return ", ".join(sorted(set(numbers), key=int))
|
| 86 |
+
|
| 87 |
def sum_food_sales(self, filepath):
|
| 88 |
df = pd.read_excel(filepath)
|
| 89 |
food_df = df[df["Category"].str.lower() == "food"]
|
|
|
|
| 104 |
if filepath := getattr(file, "name", None):
|
| 105 |
if filepath.endswith(".mp3"):
|
| 106 |
transcript = self.transcribe_audio(filepath)
|
| 107 |
+
if "ingredient" in q:
|
| 108 |
+
return self.extract_ingredients(transcript)
|
| 109 |
+
if "page" in q:
|
| 110 |
+
return self.extract_page_numbers(transcript)
|
| 111 |
return transcript
|
| 112 |
elif filepath.endswith(".xlsx") or filepath.endswith(".xls"):
|
| 113 |
return self.sum_food_sales(filepath)
|
|
|
|
| 127 |
return "An error occurred processing the question."
|
| 128 |
|
| 129 |
|
| 130 |
+
|
| 131 |
def run_and_submit_all( profile: gr.OAuthProfile | None):
|
| 132 |
"""
|
| 133 |
Fetches all questions, runs the BasicAgent on them, submits all answers,
|